Background
The primary objective of this review is to assess the accuracy of machine learning methods in their application of triaging the acuity of patients presenting in the Emergency Care System (ECS). The population are patients that have contacted the ambulance service or turned up at the Emergency Department. The index test is a machine-learning algorithm that aims to stratify the acuity of incoming patients at initial triage. This is in comparison to either an existing decision support tool, clinical opinion or in the absence of these, no comparator. The outcome of this review is the calibration, discrimination and classification statistics.
Methods
Only derivation studies (with or without internal validation) were included. MEDLINE, CINAHL, PubMed and the grey literature were searched on the 14th December 2019. Risk of bias was assessed using the PROBAST tool and data was extracted using the CHARMS checklist. Discrimination (C-statistic) was a commonly reported model performance measure and therefore these statistics were represented as a range within each machine learning method. The majority of studies had poorly reported outcomes and thus a narrative synthesis of results was performed.
Results
There was a total of 92 models (from 25 studies) included in the review. There were two main triage outcomes: hospitalisation (56 models), and critical care need (25 models). For hospitalisation, neural networks and tree-based methods both had a median C-statistic of 0.81 (IQR 0.80-0.84, 0.79-0.82). Logistic regression had a median C-statistic of 0.80 (0.74-0.83). For critical care need, neural networks had a median C-statistic of 0.89 (0.86-0.91), tree based 0.85 (0.84-0.88), and logistic regression 0.83 (0.79-0.84).
Conclusions
Machine-learning methods appear accurate in triaging undifferentiated patients entering the Emergency Care System. There was no clear benefit of using one technique over another; however, models derived by logistic regression were more transparent in reporting model performance. Future studies should adhere to reporting guidelines and use these at the protocol design stage.
Registration and funding
This systematic review is registered on the International prospective register of systematic reviews (PROSPERO) and can be accessed online at the following URL: https://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42020168696
This study was funded by the NIHR as part of a Clinical Doctoral Research Fellowship.
ObjectivesAn NHS England report highlighted key issues in how patients were initially navigating access to healthcare. This has manifested in increased pressure on ambulance services and emergency departments (EDs) to provide high quality, safe and efficient services to manage this demand. This study aims to identify non-urgent conveyances by ambulance services to the ED that would be suitable for care at scene or an alternative response.DesignA retrospective analysis of emergency department data linked to initial pre-hospital call data (either ‘111’ or ‘999’) in 2014 in Yorkshire and Humber. A previously validated definition of non-urgent attendance at ED was adapted for pre-hospital use to identify all linked ambulance conveyances that had no ‘in-hospital’ specific investigations, treatments or follow up care during that episode. Linked data was used to identify clinical triage conditions at the time and source of call (999 or 111).SettingAll 14 acute trusts (acute hospital and ambulance service) in Yorkshire and Humber.Outcome and measuresThe number of non-urgent attendances to ED which were conveyed by ambulance was examined in terms of age, time of arrival, initial triage (AMPDS) and final ED diagnosis.Figure 1ED final diagnosis for avoidable conveyances (n=65,360)Results1,312,539 linked patient episodes were analysed which included ambulance service contact and hospital data. 4 04 348 (30.8%) of the total reported were transported by ambulance. Of all the linked conveyances, 65 360 (16.2%) were classed as non-urgent ED attendances. There were significantly increased odds of a non-urgent conveyance out of hours (OR:1.48; 95% CI:1.45 to 1.51). Of all conveyances of patients aged 16–34 (n=77,683), 24 443 (31.5%) patients were non-urgent. This compares with patients aged 75+ (n=150,668), in which 11 400 (7.1%) were considered non-urgent. 70.6% of the data was included for AMPDS analysis. This demonstrated the largest numbers of non-urgent conveyances (by code) came from falls (n=5277, 8.1%) and outside referrals such as ‘Healthcare Professionals’ (n=3983, 6.1%) and the ‘111 telephone service’ (n=9437, 14.4%). ED diagnosis analysis showed the highest proportion of patients were attending with minor injury and illness, and alcohol intoxication.Conclusions16% of ambulance conveyances to ED in 2014 were non-urgent with around 1 in 3 patients under the age of 34 conveyed with non-urgent complaints. 1 in 5 patients had a non-urgent conveyance out of hours. AMPDS analysis identified target areas for intervention including referrals from other healthcare providers. Final ED diagnosis identified specific patient target areas including minor illness and alcohol intoxication.Figure 2Figure 3Age of patients taken to ED by ambulance (avoidable)
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.